SessionExprFiles

Home | Expression Sessions

Registered Files for Expression Sessions

Three main files are required, each of the following pages has detailed specifications of each file:

1. Series info: series_info.csv
2. Sample metadata: samples.csv
3. Expression data: expr.txt

These files are optional, or can be made via performing score calculations:

• Score info: score_info.csv
• Coords: coord_XYZ.csv

File names

The above files have a "GEO" prefix if they are in a GEO type session, and an "expr" prefix if they are in a base expression session. For example, the expression matrix is called "GEO.expr.txt" in a GEO session or "expr.expr.txt" in an expression sessions. Similarly, it would be "GEO.samples.csv" or "expr.samples.csv"

Troubleshooting

Read these above pages, most problems occur when formatting those first 3 files.

• I haven't worked out all the issues with using non-ASCII unicode characters and getting them through the whole workflow. I often see these characters in GEO descriptions, so you may need to edit the series info file and look for characters like: long dash, i with two dots in naive, greek characters, +/-, etc, and change them to ASCII characters.
• as stated in here, the expression matrix file (a tab delimited file) should not have a column name for the row names (which are gene names). In other words, be sure that the first character in the expr.txt file is a tab.

You can try running this R script to validate some of the potential errors in your files:

ser.info = read.csv(file="series_info.csv")
samples = read.csv(file="samples.csv")
expr = read.table(file="expr.txt", sep="\t", header=T, quote="")

if (samples[1,1] == "_id") {
  if (any(duplicated(samples[,1]))) stop("First column in samples.csv file has duplicate values")
  rownames(samples) = samples[,1]
  samples[,1] = NULL
}

expr_first_line = read.table(file="expr.txt", sep="\t", header=F, quote="", nrow=1)
if (length(expr_first_line) == ncol(expr)) stop("header row of expr.txt should only have colnames for data columns, no col name for the rownames")

if (any(duplicated(toupper(rownames(expr))))) stop("duplicate gene names")
if (! all(rownames(samples) == colnames(expr))) stop("samples ids don't match")
if (! all(colnames(ser.info) == c("key", "value"))) stop("samples ids don't match")